Although deep neural networks (DNNs) have achieved impressive performance in several non-trivial machine learning (ML) tasks, many challenges remain. One class of challenges has to do with understanding how and why DNNs obtain their improved performance, given that they do not exhibit properties such as convexity that are common among ML methods. A second class of challenges has to do with using this understanding to develop DNN methods that have better statistical/inferential properties and/or better algorithmic/running time properties. In this work, the team will pursue two related directions. First, to develop a model to make theoretically precise certain intuitions that might explain the performance of DNNs.

The AI startup landscape report has been divided into two sub-sectors -- infrastructure companies, which provide cloud infrastructure, algorithms, libraries for creating AI based applications. Examples of infrastructure services include AI as a service (Data Robot, Sentient), NLP as a service (Nuance, SpinVox), and computer vision as a service (Face). The second sub-sector -- AI applications looks at companies developing applications based on artificial intelligence technologies for different verticals. These include companies leverage AI techniques to build applications tailored for end use in enterprise (Palantir, Dataminr, Cylance), industry (ButterFly Network, Zest Finance) and consumer (X.ai, Anki) markets. Over 1.2B has been invested in AI infrastructure startups since 2010 with 540M being invested in 2015-16.

A clever hack by Alexander Crosson and Naveen Kulandaivelu at today's TechCrunch Disrupt SF Hackathon may be giving us tech journalists early warning signs of our forthcoming replacement by AI. While nowhere near advanced enough to conduct independent due diligence or investigative reporting, their project, CrunchLetter, is aimed squarely at automating the way we get news. Newsletters from publications like TechCrunch and StrictlyVC, as well as platforms like CrunchBase, are painstakingly assembled by hand, but the team's machine learning experiment leveraging Google's Tensor Flow is already able to generate rudimentary venture capital deal summaries. The tool uses unsupervised machine learning to analyze the CrunchBase data set, alongside articles from major VC publications, to generate newsletters for folks following the startup funding beat. During the team's presentation, the two showed a work in progress that could assemble a 30 word summary of a funding round.

When you're a firefighter, your life is always on the brink of danger. Now, NASA has developed an artificial intelligence system called AUDREY that could help save the lives of firefighters by nagivating them in smoke-filled burning establishments. Meet AUDREY (Assistant for Understanding Data through Reasoning, Extraction, and sYnthesis), an artificial intelligence system being developed by NASA's Jet Propulsion Laboratory. According to a report from Smithsonian, AUDREY is a technology that tracks the firefighters movements and guides them to safety during rescue. The technology originated from the Jet Propulsion Laboratory's work on space rovers kused on Mars.

The session is comprised of a panel discussion with leading researchers debating the current scientific trends in AI and its applications. That is followed by a broader discussion that dives into how the developments in AI affect our lives and society. The Heidelberg Laureate Forum Foundation (HLFF) is driven to foster the opportunity for progressive discourse, and the Hot Topic session is a crucial component of that goal. Today, AI is no longer a brash, cryptic concept taken directly from the pages of science fiction. The developments owed to the technology based on AI have altered what we thought possible and has done so in a much quicker fashion than was predicted.

For the book, I interviewed executives at more than 50 work settings across industries about how automation technologies (machine learning, robotics, wearables etc.) are transforming jobs. The practitioners I interviewed talked about using technology to improve productivity and product quality. They were pragmatic and generally optimistic. I also found a contrasting sense of pessimism, indeed panic, in the academic, analyst and economist world about machines causing'jobless futures." Why are so many smart people so pessimistic? Being a technology enthusiast for decades, I wondered what I might be missing. I found inspiration in something I had heard from Bill Joy, one of the co-founders of Sun Microsystems. Joy, who was once described by Fortune as the "Edison of the Internet," had this guidance: "If you cannot solve a problem, make the problem bigger.

Rana el Kaliouby will present a keynote, Why AI needs emotion, at the O'Reilly Artificial Intelligence Conference, September 26-27, 2016, in New York. Subscribe to the O'Reilly Data Show Podcast to explore the opportunities and techniques driving big data, data science, and AI. Find us on Stitcher, TuneIn, iTunes, SoundCloud, RSS. While I was in Beijing for Strata Hadoop World, several people reminded me of the chatbot Xiaoice--one of the most popular accounts on the Chinese social media site Weibo. Developed by Microsoft researchers, Xiaoice comes with a personality and is able to engage users in extended conversations on Weibo.

One day the AIs are going to look back on us the same way we look at fossil skeletons on the plains of Africa. An upright ape living in dust with crude language and tools, all set for extinction. Ex Machina, a Hollywood blockbuster made on a 15 million budget, tells the story of a programmer who is invited by his employer, the eccentric billionaire Nathan Bateman who built a fictional search engine called Blue Book, to administer the Turing test to an android with artificial intelligence, which essentially determines whether a computer can trick a human into believing she is having a conversation with another human. Everything in our online life is indexed. Every idle tweet, status update, or curious search query feeds the Google database.

Macquarie Bank is hoping to differentiate itself from rivals in the fiercely competitive digital banking space by using machine learning and natural language processing tools to give its customers detailed insight into their finances. The bank last week unveiled its new online and mobile banking platform for personal banking customers, a vastly significant upgrade from the online-only platform customers have been using to date. It's quite a few years late to market with such an offering, but decided to forego the competitive advantage that comes with being a first mover to wait until the stars aligned with its core banking overhaul and until it could deliver what it considers a truly differentiating offer. In designing the new platform - which consists of native iOS and Android apps and a website - the bank said it looked to customer experiences delivered by the likes of Netflix, Facebook and Spotify to create an offering that would compete with "the last app a customer used". As a result, Macquarie Bank personal banking customers can now use tools underpinned by machine learning and natural language processing technologies to tag and track transactions, automatically categorise spending, and "search-how-you-speak", among other things.

Optunity is a free software package dedicated to hyperparameter optimization to automatically find suitable hyperparameters for a given learning task. Optunity's dedicated optimizers are a drop-in replacement for grid search, which will yield better hyperparameters while requiring less computation time. The design focuses on ease of use, flexibility, code clarity and interoperability with existing software in popular machine learning environments, such as scikit-learn, OpenCV and Spark's MLlib.